How Much Grid Capacity Can 21 GWh Buy?

A large contract portfolio, a small energy footprint.

The Dutch electricity grid can be full even when it has considerable unused capacity for most of the year.

Netbeheer Nederland’s latest half-year implementation figures describe 769 capacity-steering contracts at the end of June 2026, representing approximately 4.4 GW of flexible power. A detailed transcription of the report records 21,329 MWh, or 21.3 GWh, of electricity adjustments during the first six months. Meanwhile, regional network operators still had 15,498 requests for additional withdrawal capacity, together seeking 9,986 MW. [1][2]

It is tempting to divide 21.3 GWh by 4.4 GW and conclude that the contracts were barely used. The quotient is about 4.8 hours, but that is not a utilisation measure: the power figure is a snapshot on 30 June, while the energy figure accumulates adjustments over six months. Contracts had different start dates, locations and operating conditions. The totals also combine adjustments to electricity injection with adjustments to withdrawal, which need not relieve the same network constraint.

The more useful question is different. How much additional reliable grid capacity can a network release by adjusting relatively little electricity at exactly the right time?

The hour that determines the size of the network

Consider an illustrative transformer that can safely carry 100 MW. Its existing load is lower than that limit for most of the year, but it approaches 100 MW during a short cluster of high-demand hours.

A new 5 MW factory connection may be impossible because the transformer could carry 105 MW during its busiest hour. Yet the same factory could operate without difficulty during thousands of quieter hours.

There are two ways to address the constraint. The network owner could reinforce the asset, or an existing user could agree to reduce demand during the critical period. In the second case, the benefit is the connection made possible, not the electricity saved.

For example, cutting demand by 5 MW for 50 hours adjusts only 250 MWh. If those are all the hours that matter, the intervention can support a much larger volume of electricity use over the rest of the year. But if another constrained hour falls outside the contract, the network may still be unable to promise the full new connection.

That last condition is the heart of the economics. It is not enough for flexible MW to exist somewhere in the country. They must reduce the flow across the particular constrained asset, in the correct direction, when it would otherwise be overloaded.

A model shows where the value stops rising

To test the mechanism, we constructed a simple local network rather than claiming that the national figures describe one. Our synthetic 100 MW transformer has 200 ranked peak hours, with existing load falling gradually from 100 to 90 MW. A hypothetical 5 MW flexibility contract can be delivered only in a selected number of the highest-load hours.

In this deliberately idealised calculation, the extra firm capacity a new customer can use changes sharply with the number of critical hours covered.

Illustrative calculations only. All inputs are constructed; the figures are not observed connection releases in the Netherlands.

At 50 covered hours, the next uncovered high-load hour is still too busy to accommodate the full 5 MW. At 100 hours, the whole target becomes feasible. Another 50 hours of potential availability add no further firm capacity in this scenario.

The lesson is not that 100 hours is a universal threshold. The threshold depends on the local load shape, the requested connection and the amount of flexibility that can actually be delivered. What matters is that the benefit can rise steeply and then stop, even while the contract remains available for more hours.

Is flexibility cheaper than another transformer?

A small energy adjustment does not automatically imply a cheap solution. Businesses may need to reschedule production, maintain backup capacity or accept operating risk. Network operators may also pay to reserve flexibility whether or not they call on it.

We tested an explicitly hypothetical investment choice. Assume the full 5 MW connection can be supplied either by reinforcing the network now at a cost of €4 million, or by using secure flexibility for three years and constructing the same upgrade later. At a 5% real discount rate, postponing €4 million for three years creates €545,000 of present-value timing benefit.

Now assume the flexibility contract costs €20,000 per MW-year for availability, plus €100 for each MWh adjusted. With the model’s 251 MWh annual adjustment, those payments have a three-year present value of approximately €341,000. The remaining illustrative operator advantage is around €204,000.

This is not a measured Dutch saving or a quoted flexibility tariff. It is an auditable comparison of two ways to provide the same 5 MW service. It is valid only if the full connection is genuinely feasible during the waiting period and the €4 million upgrade really can be postponed. Deferring an asset is not the same as avoiding its capital cost forever.

Illustrative operator comparison: a longer deferral raises the cumulative present-value advantage when annual contract payments remain below the upgrade timing benefit. This is not a Dutch tariff or investment return.

There is another possible benefit when new customers otherwise have to wait: earlier production, operations or construction. That needs a different comparison, based on additional net economic value from earlier access, not a multiplication of grid MW by company turnover. It should not be added mechanically to the operator’s financial saving.

The missing metric

The national publication reports the number and size of contracts, how often they were activated and how much electricity was adjusted. It does not establish how many new firm MW were caused by those activations, or which reinforcement projects were deferred because of them.

To answer that, each congestion area needs its own ledger: the binding asset, the critical hours, the flexibility delivered at those hours, the additional connection actually approved, the alternative reinforcement and the cost of the whole arrangement.

That information also permits an important negative finding. A contract that adjusts plenty of electricity may be worth little for new connections if it is located on the wrong side of a bottleneck. A rarely activated contract may be extremely valuable if it provides reliable insurance against the one remaining critical hour. Neither outcome can be read from national MWh totals.

Dutch operators and the regulator are already working to improve flexible contracts and network-use transparency. ACM’s June 2026 agreements explicitly called for better insight into network use and wider implementation of congestion-management arrangements. [3]

Capacity is not an annual energy total

The distinction matters beyond the Netherlands. Factories, data centres, charging networks, housing projects and electricity producers make investment decisions based on whether an actual connection is available, not the annual average loading of the grid.

The relevant economic question for infrastructure owners is therefore not how much energy was shifted? It is how many additional, secure megawatts of network service were made possible, at what total cost, and for how long?

In areas where only a narrow peak binds, targeted flexibility may postpone substantial construction or accelerate productive investment. Where the network is constrained across many hours or physical bottlenecks, building remains necessary. The Dutch data show that the flexibility market is expanding; only local counterfactual evidence can establish how much usable infrastructure capacity it is creating.

Research note

The national figures are reported observations, with detailed activation statistics taken from a published transcription of the operator report pending direct PDF verification. The network and financial results are hypothetical scenarios, not Dutch feeder estimates. The workbook includes input registers, coverage sensitivity, separate financial and welfare comparisons, and formula checks.

Explore the research

The technical report derives the capacity constraint, explains why contracted MW cannot be treated as released firm MW and tests when flexibility is cheaper than deferring reinforcement. The accompanying Excel model lets readers change the hypothetical peak shape, hours of coverage, delivered power and contract prices, with each assumption visibly labelled.

Sources

[1] Netbeheer Nederland, Slimmer netgebruik wint terrein, maar vraag naar transportcapaciteit groeit nog altijd sneller, 5 October 2026. https://www.netbeheernederland.nl/artikelen/nieuws/slimmer-netgebruik-wint-terrein-maar-vraag-naar-transportcapaciteit-groeit-nog

[2] Solar Magazine, Bedrijven zetten massaal in op flexibel stroomverbruik om netcongestie te verlichten, 5 October 2026, transcribing the report’s detailed capacity-steering figures. https://solarmagazine.nl/nieuws-zonne-energie/i44986/bedrijven-zetten-massaal-in-op-flexibel-stroomverbruik-om-netcongestie-te-verlichten-4-4-gigawatt-flexvermogen-gecontracteerd

[3] Autoriteit Consument & Markt, ACM en netbeheerders maken bindende afspraken, 2 June 2026. https://www.acm.nl/nl/publicaties/acm-en-netbeheerders-maken-bindende-afspraken-over-aanbieden-flexibele-contracten-inzet-congestiemanagement-en-beter-netinzicht